Waveguides can transmit very high powers, isolating the signal from external noises and interferences. Besides, waveguides have a very low loss. These capabilities make them an interesting candidate for signal transmission between two cities. Why are rectangular waveguides not used for inter-city transmission?
I guess it may be because rectangular waveguides have a narrow bandwidth, and therefore it is necessary to use many of them for signal transmission which is impractical. Am I right?
The medium inside of a waveguide is occupied by gas. It could be vacuum, probably even with less loss. However, what should not be in there is water. It is nearly impossible to prevent water in the miles and 10's of thousands of joints necessary for waveguides.
Optical waveguides, i.e. fiber, are solid, and therefore prevent the intrusion of water instantaneously, and somewhat on long term basis as well. Granted, glass fiber and its jacket WILL absorb 'microscopic' amounts of water, causing high loss. But it takes awhile and is easy to prevent with a very tiny amount of material on each joint. It's also highly effective sealing.
Undersea fiber optic links are amazing. Every so often a fiber optic amplifier, made of fiber, is insterted in series. The energy for the fiber optic laser is ANOTHER laser shooting all the way to the other continent. Using splitters and combiners, a small amount of the LOWER frequency (longer wavelength) power laser is sent through a specially doped piece of fiber, keeping the dopant atoms in an excited state. As the pulsed signal laser combines in the laser amplifier fiber, it triggers additional laster power from the exicted atoms in the amplifier and well, amplification happens :-)
Another part of the puzzle is called time dispersion. Not all photons take the exact same path in the fiber. Some hug and bounce off the walls, some go down the center. So not all arrive at the same time, since having traveled microscopically different path lengths. This cause the amplitude of the energy delivered by the photons to be spread out, the wave form does NOT instantaneously jump to full amplitude. This limits the bandwidth the longer the fiber.
The ingenious physicists and optical engineers figured out if the made fiber where the lightspeed is slower in the center to than at the outer wall in a glass fiber, that the photons could all be realigned in time as exit this 'correction fiber'. Since they made the change in speed significant, it only takes a small amount of fiber every kilometer or so to make the correction.
NOW, all of this is built into a cable assembly, sealed, and dropped into the ocean. The asembly is done on a ship at sea as they drop it, or in a truck on the side of the trench on land. I've watched some of it being done on land. Amazing. The most amazing part is, there is no electricity or electronics in the entire cable for THOUSANDS OF MILES. All of the reamplification and waveform reshaping happends optically as described above. I forgot to mention that since the power laser is lower wavelength and continuous wave, it has a very low loss in the fiber, and can go to at least the halfway point. They could then inject power laser from the OTHER continent to the midway point to amplify the signals the rest of the way to the target continent.
NONE OF THIS is possible in the RF domain. And as others said, the bandwith is insane. Nowadays, they can add channels via: wavlength discrimination, polarisation discrimination, optical rotation along the center axis, and spirally injected light in aa spiraled do-nut shape down the fiber. Quite a few others are being attempted. So fiber bandwidth will continue to climb for a while, using fibers already installed!
Waveguides over several miles would be wildly expensive and unstable. How would you hold up miles of expensive precision-machined pipe? It would sag under its own weight. Temperature changes would make it challenging to design. There's the need for raw material per mile to make such waveguides, and maintenance per mile per year.
Open air costs zero per mile, and takes no maintenance between endpoints except occasional tree trimming, so EM radiation wins the economics contest. All the expense goes into antenna design and fabrication, including short-run waveguides, at each endpoint, not vast quantities of material between points. That scales better when building a nation-scale network.
Waveguides were actually used for a short time, the Bell System developed a network based on round underground waveguides and even built a pilot factory.
Here is a short brochure http://long-lines.net/tech-equip/radio/WE-waveguide/WEWP-1.html and an article https://archive.org/details/bstj43-4-1783
Partly because of this investment, they were a few years late transitioning to optical waveguides, which are much cheaper and have much higher bandwidth.
Plenty of technical details can be found in the book "A History of Engineering and Science in the Bell System: Transmission Technology (1925-1975)", a popular account in "The Idea Factory" by Gertner. Both are great books.
There are multiple reasons why this is never done:
The main advantage of using RF is you can transmit it through space relatively robustly. Putting it in a waveguide loses this advantage.
Waveguides are made out of metal and building very long, precision waveguides and then installing them in the ground or hanging them on poles is extremely expensive. On top of that, RF in general (in a waveguide or in free space) is more or less limited to under 100 GHz of bandwidth.
On the other hand, optical fiber is just glass and so is quite cheap. Optical fiber is also one of the most low loss materials around - good transmission grade fiber can have a loss of around 0.2 dB per km. Yes, you only lose 20 dB when you go through 100 km of fiber, and it's very easy to boost that back up with fiber amplifiers at regular intervals.
Fiber also provides an absolutely huge bandwidth and it is immune to external EM interference. It's trivial (though not so cheap) to put 100 or more signals through a single fiber on 100 GHz or 50 GHz centers and move several Tbps.
It is even possible to modulate analog RF onto laser light (with several GHz of bandwidth) and transmit that down a fiber, possibly even with multiple of these channels in parallel. This is called RF over fiber, and it is used occasionally for things like connecting broadcast stations to transmitters.
The bandwidth through a fiber is absolutely huge because the center frequency is in the 100s of THz. RF doesn't get anywhere near that.
The BT Trunked Waveguide trial was an effort to use high capacity waveguide (300,000 voice calls) on telephone trunk routes - it was state-of-the-art technology in its day. The waveguide was actually circular, copper wire was spun on a mandrel to make a tube. It was probably easier to make than rectangular waveguide but was still expensive - copper, expensive to install - trenching near straight lines, and expensive to maintain - keeping it pressurized to keep out moisture, (another reason rectangular cross section is not preferred) etc.
Then fibre optics came along and made trunked waveguide redundant. The installed copper was so valuable it was economically viable to rip up the trial waveguide for scrap.
More here in Short History of Telecommunications Transmission in the UK: pp37
I arrived at BT Research Labs some years after this project was cancelled. It was still talked about as proof of why you have to invest in researching different technologies... one of them might render everything else obsolete.
